Br. J. Cancer (1 976) 34, 227
CARCINOEMBRYONIC ANTIGEN AND GLUCOSE PHOSPHATE ISOMERASE IN A HUMAN COLONIC CANCER MODEL (GW-39) D. MUNJAL AND D. M. GOLDENBERG
From the Division of Experimental Pathology, Department of Pathology, University of Kentucky Medical Center, Lexington, Kentucky 40506 Received 6 April 1976
Accepted 20 May 1976
Summary.-Levels of carcinoembryonic antigen (CEA) and glucose phosphate isomerase (GPI) have been compared in the circulating blood of hamsters bearing intra-muscular grafts of GW-39 human colonic tumour. CEA in the sera of GW-39 tumour-bearing hamsters ranged from 2 6 to 8-4 ng/ml (mean = 4-5 ± 1-7 ng/ml). GPI in the sera of normal hamsters ranged from 332 to 749 iu/l (mean = 602 ± 110 iu/l) while those with 14-week-old intra-muscular grafts of a hamster amelanotic melanoma, (A.Mel.3), or GW-39 human colonic carcinoma had a range of 664 to 1267 iu/l (mean = 1024 + 220 iu/l) and 1430 to 4719 iu/l (mean - 2065 ± 601 iu/l) respectively. Thus, the ratio of enzyme activity in GW-39, A.Mel.3, and normal hamsters was 3 4:1-7:1, indicating a significant elevation (P < 0.01) in animals bearing a human colon carcinoma or a hamster melanoma, with particularly high values obtained in hamsters with GW-39. Sequential determinations of CEA and GPI in a group of hamsters transplanted intra-muscularly with GW-39 tumours revealed that both markers increased proportionately with duration of tumour growth, suggesting that both serum CEA and GPI may be used as measures of tumour growth. The concentration of GPI in GW-39 human colonic carcinoma xenografts was also significantly higher than that measured in normal human colon, primary human colonic cancer, or normal hamster tissues. These results support the view that GPI, in addition to CEA, is a quantitatively increased marker in this tumour model, and is liberated into the circulation in proportion to the increase in tumour mass.
TUMOUR ANTIGENS rarely, if ever, have been found to be truly tumour- or organspecific (Laurence and Neville, 1972). The carcinoembryonic antigen (CEA) of Gold and Freedman (1965), although originally considered to be specific for digestive tract cancers, has likewise not realized its potential as a specific diagnostic test for this cancer type (Hansen et al., 1974; Zamcheck, 1974). Nevertheless, it has proved to be an important stimulus in the search for other tumour markers. Since earlier studies have indicated that certain serum enzymes can be significantly elevated in cancer patients (Badenetal., 1971; Bodansky, 1954, 1974), it was considered of interest to assess the combined use of CEA and particular
serum enzymes as a possible improvement over each modality by itself in the detection of cancer. Recent studies of cancer patients have indeed shown that some of these enzymes can increase diagnostic accuracy in breast, lung and colorectal cancers when combined with the plasma CEA test (Steele et al., 1974; Cooper et al., 1975; Mun.jal et al., 1976). As a corollary to these initial clinical studies, we have undertaken an evaluation of the circulating levels of CEA and GPI in hamsters bearing a xenografted human colonic carcinoma, GW-39 (Goldenberg, Witte and Elster, 1966), in order to study certain relationships of these two putative tumour markers in a human tumour model.
228
D. MUNJAL AND D. M. GOLDENBERG
reagents supplied by Hoffmann-La Roche, Inc., Nutley, N.J. Appropriate standard Tumours.-GW-39 tumours were pro- curves were made with normal hamster sera pagated in the hind limb musculature of and the samples were diluted in 0.9% NaCl unconditioned adult hamsters (Sprague- before extraction with equal volumes of Dawley, Madison, Wisc.) of both sexes, 1 2M percholoric acid. Due to the relatively weighing 60 to 80g. The tumours were larger serum quantity required for the CEA excised at regular intervals between 6 and assay, it was necessary to pool the serum 16 weeks. The tumour transplantation samples. GPI activity in the serum (50 ,ul/specitechnique is that of Goldenberg et al. (1966), and consists of injecting 0-1 ml of a 10 to 20% men) was measured by the method of (w/v) tumour cell suspension into the growth Bueding and MacKinnon (1955), using stansite. Expansively-growing, mucin-producing, dard reagents supplied by Worthington signet-ring-cell carcinomas result, in almost Biochemical Corporation, Freehold, N.J. all animals grafted. Ten hamsters bearing Each animal's serum specimen was indiFortner's hamster amelanotic melanoma, vidually assayed for GPI activity. GPI A.Mel.3 (Fortner, Mahy and Schrodt, 1961), catalyzes the isomerization of fructose-6at the same growth site, as well as 25 un- phosphate to glucose-6-phosphate, which is treated hamsters, served as controls. Another in turn oxidized to 6-phosphogluconate and NADH in the presence of glucose-6-phosphate group of 5 hamsters bearing intra-muscular (i.m.) grafts of GW-39 was sequentially bled dehydrogenase and NAD. The NADH by cardiac puncture up to 16 weeks after produced in the second part of the reaction is directly proportional to the glucose-6-phostransplantation. Sample collection.-The blood was collected phate produced in the first part of the by cardiac puncture and the serum was reaction, so that the rate of increase of separated. The same serum sample was used NADH measured at 340 nm is a measure of for CEA and GPI tests. During bleeding, GPI activity. One international unit (iu) reduces 1 ,umol NAD per min at 30°C. care was taken to avoid haemolysis since Protein.-Protein was quantitated by the some blood components release, inhibit, or activate the enzyme. Plasma samples were Lowry et al. (1951) procedure, using bovine avoided, since variable results have been serum albumin as the reference standard. reported irrespective of the anticoagulant The concentration of GPI in tissue extracts was calculated as iu/g protein. used (Harrocks, Ward and King, 1963). Tumour and normal tissue extractions.Tissues (lung, liver, colon, spleen and kidney) from normal hamsters or GW-39 tumours RESULTS from tumour-bearing hamsters were collected by sacrificing the animals. Normal and The results for GPI activity and CEA malignant human tissue specimens were in the sera of normal hamsters and obtained at surgery or autopsy. The tissues hamsters bearing i.m. grafts of A.Mel.3 or were washed free of blood components with ice-cold distilled water and necrotic parts, if GW-39 tumours are presented in Table I. GPI activity in the sera of normal hamany, were dissected away. Pooled normal tissues or GW-39 tumours were minced and sters ranged from 332 to 749 iu/l, with a homogenized in 5 volumes (w/v) cold distilled mean of 602 + 100 iu/l. Hamsters bearH20 in a Sorvall Omnimixer. Following ing 14-week-old i.m. grafts of A.Mel.3 and centrifugation for 30 min at 10,000 rev/min GW-39 human colonic carcinoma had at 40C, the pellets were rehomogenized in 3 serum GPI values of 664 to 1267 iu/l (with volumes of cold distilled H20 and recentri- a mean of 1024 + 200 iu/l) and 1430 to fuged. The combined supernates were tested 4719 iu/l (with a mean of 2065 ± 601 iu/l), for CEA, GPI and protein content. respectively. Thus, the ratio of GPI in Measurement of CEA and GPI levels.CEA in the serum (0-5 ml/specimen) was the circulating blood of GW-39 animals, measured by an indirect radioimmunoassay A.Mel.3 animals, and normal hamsters is using Hansen's Z-gel procedure (Hansen, 3-4:1:7: 1. Levels of CEA in the sera of Lance and Krupey, 1971), with standard hamsters bearing GW-39 tumours sacriMATERIALS AND METHODS
G~ PI
229
CEA AND GPI IN HUMAN COLONIC CAXOER
TABLE I.-Levels of GPI and CEA in Circulating Blood of Tumour-bearing and Normal Hamsters No. of animals 25 10 65
CEAt (ng/ml) GPI (iu/l)
No. of , animals 10
A_
Group Range P Range Mean±s.d. Mean± s.d. Normal 332-749 0 3 0-0 5 602±110 A.Mel.3 664-1267
CARCINOEMBRYONIC ANTIGEN AND GLUCOSE PHOSPHATE ISOMERASE IN A HUMAN COLONIC CANCER MODEL (GW-39) D. MUNJAL AND D. M. GOLDENBERG
From the Division of Experimental Pathology, Department of Pathology, University of Kentucky Medical Center, Lexington, Kentucky 40506 Received 6 April 1976
Accepted 20 May 1976
Summary.-Levels of carcinoembryonic antigen (CEA) and glucose phosphate isomerase (GPI) have been compared in the circulating blood of hamsters bearing intra-muscular grafts of GW-39 human colonic tumour. CEA in the sera of GW-39 tumour-bearing hamsters ranged from 2 6 to 8-4 ng/ml (mean = 4-5 ± 1-7 ng/ml). GPI in the sera of normal hamsters ranged from 332 to 749 iu/l (mean = 602 ± 110 iu/l) while those with 14-week-old intra-muscular grafts of a hamster amelanotic melanoma, (A.Mel.3), or GW-39 human colonic carcinoma had a range of 664 to 1267 iu/l (mean = 1024 + 220 iu/l) and 1430 to 4719 iu/l (mean - 2065 ± 601 iu/l) respectively. Thus, the ratio of enzyme activity in GW-39, A.Mel.3, and normal hamsters was 3 4:1-7:1, indicating a significant elevation (P < 0.01) in animals bearing a human colon carcinoma or a hamster melanoma, with particularly high values obtained in hamsters with GW-39. Sequential determinations of CEA and GPI in a group of hamsters transplanted intra-muscularly with GW-39 tumours revealed that both markers increased proportionately with duration of tumour growth, suggesting that both serum CEA and GPI may be used as measures of tumour growth. The concentration of GPI in GW-39 human colonic carcinoma xenografts was also significantly higher than that measured in normal human colon, primary human colonic cancer, or normal hamster tissues. These results support the view that GPI, in addition to CEA, is a quantitatively increased marker in this tumour model, and is liberated into the circulation in proportion to the increase in tumour mass.
TUMOUR ANTIGENS rarely, if ever, have been found to be truly tumour- or organspecific (Laurence and Neville, 1972). The carcinoembryonic antigen (CEA) of Gold and Freedman (1965), although originally considered to be specific for digestive tract cancers, has likewise not realized its potential as a specific diagnostic test for this cancer type (Hansen et al., 1974; Zamcheck, 1974). Nevertheless, it has proved to be an important stimulus in the search for other tumour markers. Since earlier studies have indicated that certain serum enzymes can be significantly elevated in cancer patients (Badenetal., 1971; Bodansky, 1954, 1974), it was considered of interest to assess the combined use of CEA and particular
serum enzymes as a possible improvement over each modality by itself in the detection of cancer. Recent studies of cancer patients have indeed shown that some of these enzymes can increase diagnostic accuracy in breast, lung and colorectal cancers when combined with the plasma CEA test (Steele et al., 1974; Cooper et al., 1975; Mun.jal et al., 1976). As a corollary to these initial clinical studies, we have undertaken an evaluation of the circulating levels of CEA and GPI in hamsters bearing a xenografted human colonic carcinoma, GW-39 (Goldenberg, Witte and Elster, 1966), in order to study certain relationships of these two putative tumour markers in a human tumour model.
228
D. MUNJAL AND D. M. GOLDENBERG
reagents supplied by Hoffmann-La Roche, Inc., Nutley, N.J. Appropriate standard Tumours.-GW-39 tumours were pro- curves were made with normal hamster sera pagated in the hind limb musculature of and the samples were diluted in 0.9% NaCl unconditioned adult hamsters (Sprague- before extraction with equal volumes of Dawley, Madison, Wisc.) of both sexes, 1 2M percholoric acid. Due to the relatively weighing 60 to 80g. The tumours were larger serum quantity required for the CEA excised at regular intervals between 6 and assay, it was necessary to pool the serum 16 weeks. The tumour transplantation samples. GPI activity in the serum (50 ,ul/specitechnique is that of Goldenberg et al. (1966), and consists of injecting 0-1 ml of a 10 to 20% men) was measured by the method of (w/v) tumour cell suspension into the growth Bueding and MacKinnon (1955), using stansite. Expansively-growing, mucin-producing, dard reagents supplied by Worthington signet-ring-cell carcinomas result, in almost Biochemical Corporation, Freehold, N.J. all animals grafted. Ten hamsters bearing Each animal's serum specimen was indiFortner's hamster amelanotic melanoma, vidually assayed for GPI activity. GPI A.Mel.3 (Fortner, Mahy and Schrodt, 1961), catalyzes the isomerization of fructose-6at the same growth site, as well as 25 un- phosphate to glucose-6-phosphate, which is treated hamsters, served as controls. Another in turn oxidized to 6-phosphogluconate and NADH in the presence of glucose-6-phosphate group of 5 hamsters bearing intra-muscular (i.m.) grafts of GW-39 was sequentially bled dehydrogenase and NAD. The NADH by cardiac puncture up to 16 weeks after produced in the second part of the reaction is directly proportional to the glucose-6-phostransplantation. Sample collection.-The blood was collected phate produced in the first part of the by cardiac puncture and the serum was reaction, so that the rate of increase of separated. The same serum sample was used NADH measured at 340 nm is a measure of for CEA and GPI tests. During bleeding, GPI activity. One international unit (iu) reduces 1 ,umol NAD per min at 30°C. care was taken to avoid haemolysis since Protein.-Protein was quantitated by the some blood components release, inhibit, or activate the enzyme. Plasma samples were Lowry et al. (1951) procedure, using bovine avoided, since variable results have been serum albumin as the reference standard. reported irrespective of the anticoagulant The concentration of GPI in tissue extracts was calculated as iu/g protein. used (Harrocks, Ward and King, 1963). Tumour and normal tissue extractions.Tissues (lung, liver, colon, spleen and kidney) from normal hamsters or GW-39 tumours RESULTS from tumour-bearing hamsters were collected by sacrificing the animals. Normal and The results for GPI activity and CEA malignant human tissue specimens were in the sera of normal hamsters and obtained at surgery or autopsy. The tissues hamsters bearing i.m. grafts of A.Mel.3 or were washed free of blood components with ice-cold distilled water and necrotic parts, if GW-39 tumours are presented in Table I. GPI activity in the sera of normal hamany, were dissected away. Pooled normal tissues or GW-39 tumours were minced and sters ranged from 332 to 749 iu/l, with a homogenized in 5 volumes (w/v) cold distilled mean of 602 + 100 iu/l. Hamsters bearH20 in a Sorvall Omnimixer. Following ing 14-week-old i.m. grafts of A.Mel.3 and centrifugation for 30 min at 10,000 rev/min GW-39 human colonic carcinoma had at 40C, the pellets were rehomogenized in 3 serum GPI values of 664 to 1267 iu/l (with volumes of cold distilled H20 and recentri- a mean of 1024 + 200 iu/l) and 1430 to fuged. The combined supernates were tested 4719 iu/l (with a mean of 2065 ± 601 iu/l), for CEA, GPI and protein content. respectively. Thus, the ratio of GPI in Measurement of CEA and GPI levels.CEA in the serum (0-5 ml/specimen) was the circulating blood of GW-39 animals, measured by an indirect radioimmunoassay A.Mel.3 animals, and normal hamsters is using Hansen's Z-gel procedure (Hansen, 3-4:1:7: 1. Levels of CEA in the sera of Lance and Krupey, 1971), with standard hamsters bearing GW-39 tumours sacriMATERIALS AND METHODS
G~ PI
229
CEA AND GPI IN HUMAN COLONIC CAXOER
TABLE I.-Levels of GPI and CEA in Circulating Blood of Tumour-bearing and Normal Hamsters No. of animals 25 10 65
CEAt (ng/ml) GPI (iu/l)
No. of , animals 10
A_
Group Range P Range Mean±s.d. Mean± s.d. Normal 332-749 0 3 0-0 5 602±110 A.Mel.3 664-1267
